The Tale of Four Rides
On its third-quarter earnings call on November 4, 2026, Uber revealed a statistic that was both comical and profound: during its busiest week in October, third-party AI agents booked a grand total of four rides. In that same period, the company processed
roughly 320 million transactions across rides and Uber Eats. Food delivery fared slightly better, with AI agents placing about 500 orders. While tech companies are heavily invested in the idea of AI 'agents' that can autonomously manage our digital tasks, these numbers paint a picture not of a revolution, but of a very quiet, tentative experiment. The data, echoed by similar reports from DoorDash and Airbnb, provides a dose of reality. The future where you simply tell your phone to 'get me a ride to the airport' and it handles the rest is not quite here.
What is an AI Agent Anyway?
An AI agent is more than just a chatbot. It's a piece of software designed to understand a goal and then take a series of actions across different apps and websites to achieve it. Instead of you opening the Uber app, typing a destination, selecting a car, and confirming, an agent could theoretically do it all for you. Companies like Meta, Google, and OpenAI are all racing to build the dominant personal AI assistant, a universal remote for your digital life. The promise is to automate tedious digital chores. However, the four-ride figure suggests a critical question: is booking an Uber really a chore that needs automating? Uber CEO Dara Khosrowshahi noted that booking a ride on the app already takes an average of only 50 seconds, raising the bar for how much time an agent could realistically save.
The Platform Pushback
While the low usage numbers might seem like a failure for AI, they also highlight the immense power held by platforms like Uber, DoorDash, and Airbnb. These companies are not just passive services waiting to be used by outside AI; they are complex ecosystems with their own data, customer relationships, and internal AI capabilities. Uber, for example, already uses sophisticated AI for demand forecasting, routing, and pricing. These companies face a strategic dilemma: if they allow third-party agents to become the main interface for customers, they risk being demoted to an interchangeable, low-margin utility. In response, they are developing their own AI tools. Airbnb is building a proprietary travel agent for 2027, believing that planning a trip is a visual, social experience that users don't want to fully delegate. Uber Eats has 'Cart Assistance' which helps fill a shopping basket from a screenshot of a recipe.
The Real Bottleneck Is Experience
The main point about the four AI-booked rides isn't that the technology has failed, but that it is being applied to the wrong problems. Booking a ride is already a solved, low-friction task for most users. The process of choosing a restaurant on DoorDash or browsing listings on Airbnb is, for many, part of the experience itself, not a burden to be offloaded. Where agents may eventually thrive is in complex, multi-step tasks that are genuinely tedious, like managing expense reports or coordinating a group trip across multiple services. The Halloween week data shows that consumers are not yet willing to trade the control and simplicity of a well-designed app for the black box of an AI agent, especially when the time saved is minimal and the potential for error is unknown.
















